Machine Learning Under Test
summary
Shows concrete testable claims and code: uses numpy.testing and unittest.mock for numeric assertions, inspects np.finfo(eps) and assert_array_almost_equal tolerances, and exercises train_test_split and ShuffleSplit to demonstrate cross validation. Generates synthetic data and fits polynomial features degree 1, 4 and 15 to reveal underfitting, a good fit at degree 4, and severe overfitting at 15, visualizes learning curves and uses confusion matrices and k‑NN examples to validate whether models actually generalize.
AI-written summary · about
chapters (21)
- 0:00Introduction
- 2:07Agenda
- 4:39Linear Regression
- 4:58Classification
- 5:26Clustering
- 5:56Supervisor Learning
- 10:40Mush Learning
- 10:53Complexity
- 12:07Risk
- 12:44Fault
- 14:57Unit Test
- 16:20Number Problems
- 21:04Newbie Testing
- 21:48MachFramer
- 22:46Function
- 23:22Output
- 25:33Model performance
- 27:48Cross validation
- 27:59Train Test Plate
- 28:39Confusion Matrix
- 31:27Recommendation
related talks
27mPyTesting the Limits of Machine Learning
42mMachine Learning Crash Course
01h 33mUnderstanding and diagnosing your machine-learning models
31mTesting Machine Learning Models
30mMachine Learning: Beyond the Tutorial
53mUnleashing the Black Box: Testing Machine Learning Systems
43mMachine Learning Testing Ecosystem of Python
33mMachine Learning explained to humans